| dc.description.abstract | The electric power grid is undergoing sustained disturbances. In particular, the extreme dynamic events disrupt normal electric power transfer, degrade power system operating conditions, and may lead to catastrophic large-scale blackouts. Accordingly, control applications are deployed to detect the inception of extreme dynamic events, and mitigate their causes appropriately, so that normal power system operating conditions can be restored. In order to achieve this, the operating conditions of the power system should be accurately characterized in terms of the electrical quantities that are crucial to control applications. Currently, the power system operating conditions are obtained through SCADA system and the synchrophasor system. Because of GPS time-synchronized waveform sampling capability and higher measurement reporting rate, synchrophasor system is more advantageous in tracking the extreme dynamic operating conditions of the power system.
In this Dissertation, a phasor parameter calculation approach is proposed to accurately characterize the power system operating conditions during the extreme electromagnetic and electromechanical dynamic events in the electric power grid. First, a framework for phasor parameter calculation during both electromagnetic and electromechanical dynamic events is proposed. The framework aims to satisfy both P-class and M-class PMU algorithm design accuracy requirements with a single algorithm. This is achieved by incorporating an adaptive event classification and algorithm model switching mechanism, followed by the phasor parameter definition and calculation tailored to each identified event. Then, a phasor estimation technique is designed for electromagnetic transient events. An ambient fundamental frequency estimator based on UKF is introduced, which is leveraged to adaptively tune the DFT-based algorithm to alleviate frequency leakage. A hybridization algorithm framework is also proposed, which further reduces the negative impact caused by decaying DC components in electromagnetic transient waveforms. Then, a phasor estimation technique for electromechanical dynamics is introduced. A novel wavelet is designed to effectively extract time-frequency features from electromechanical dynamic waveforms. These features are then used to classify input signal types, so that the PMU algorithm modeling can be thereafter tailored specifically to match the underlying signal features for the identified event. This adaptability of the proposed algorithm results in higher phasor parameter estimation accuracy. Finally, the Dissertation hypothesis is validated through experimental testing under design and application test use cases. The associated test procedures, test use cases, and test methodologies and metrics are defined and implemented. The impact of algorithm inaccuracy and communication network distortion on application performance is also demonstrated. Test results performance is then evaluated. Conclusions, Dissertation contributions, and future steps are outlined at the end. | en |
